ENHANCING EFL WRITING THROUGH PEER ASSESSMENT INTEGRATED WITH AI-BASED AUTOMATED CORRECTION: A BLENDED LEARNING INTERVENTION

Abstract

While automated writing evaluation (AWE) systems provide granular feedback on linguistic features, their capacity to address thematic depth and organizational coherence remains limited. This quasi-experimental study proposed a triadic feedback model integrating AWE (iWrite platform), structured peer assessment, and teacher scaffolding. Seventy-one Chinese university students were assigned to an experimental group (AWE + peer assessment) or control group (teacher feedback only). Pre/post-tests assessed writing proficiency using CET-4 rubrics, supplemented by Likert-scale questionnaires and textual analysis. Results indicated: (1) Significant gains in the experimental group’s overall writing scores (*d* = 1.24, *p* < .001); (2) Improved student engagement in revision cycles (97.2% implemented peer suggestions); (3) Enhanced metacognitive awareness of text organization. The intervention demonstrates how technology-mediated peer assessment can augment cognitive and social dimensions of writing development.